Exploratory study of high-strength, low-cost through-bolt with cross-pipe and nut connections for square and roundwood timber frame construction

Transcription

1 Exploratory study of high-strength, low-cost through-bolt with cross-pipe and nut connections for square and roundwood timber frame construction C. A. Eckelman Abstract An exploratory test program was undertaken to determine the withdrawal strengths of through-bolt with cross-pipe and nut connections in the ends of selected timbers. Results of the tests indicate that high strength end connections (25,000 lb.) can be obtained in structural timbers with these connectors, but surface checks and drying splits significantly reduced the strength of the joints and lead to major variations in strength. However, the weakening effect of major drying splits can be reduced by orienting the cross-pipe perpendicular to the plane of the split. Furthermore, average withdrawal strength is increased and variability in strength greatly reduced when the ends of the member are reinforced with some form of strapping or with cross bolts. Construction of joints with the nut located on the outside as opposed to the inside of the pipe provides room for the nut to be tightened in place and greatly increases possible uses of the joint since it allows direct end to end joining of members and other constructions that otherwise would not be feasible. Cost of the connectors is expected to be low $0.25 to $0.50 per kip. Through-bolt with dowel-nut connectors are commonly used in furniture where high strength mechanical connections are required. During early periods of use, dowel-nuts were often called barrel nuts (Eckelman 1977), but the term dowel-nut is more descriptive, and the author has referred to them as such in the above and all subsequent publications (Eckelman 1989, 1999; Eckelman and Erdil 2000). The high pull-out strength noted for dowel-nuts inserted cross ways through the ends of table legs led to their investigation as potential timber frame connectors by Eckelman and Senft (1995). The principal objective of that study, carried out with 6- to 7-inch- yellow-poplar peeler cores and 1-1/2-inch steel dowel-nuts with a 6-inch end spacing, was to obtain initial estimates of the ultimate tensile strengths that could be obtained with this connection (Fig. 1). Results of the tests indicated that maximum strength values are obtained when the longitudinal axis of the dowel-nut is located perpendicular to the plane of any major drying split. Strength values for with the dowel-nut located in this manner averaged 22,820 pounds. Specimens with dowel-nuts placed in the plane of the split averaged only 61 percent as strong as those placed perpendicular to the split with minimum values only 31 percent as strong. Further, it was found that strength could be increased significantly by banding the ends of the members with a 3/4-inch steel strap. Banded with the dowel-nut located in the plane of the split averaged 92 percent as strong as those with the nut located perpendicular to the plane of the split. Subsequently, the high strength of these solid steel dowel-nut connections was noted by both Wolfe (2000) and Stern (2001). Research was carried out by Wolfe et al. (2000) to determine the withdrawal strength of dowel-nuts in Douglas-fir peeler cores and small- round ponderosa pine. They found that 1.75-inch- dowelnut connectors were economically feasible at a design capacity of 10,000 pounds for 5-inch- Douglas-fir peeler cores (Wolfe et al. 2000). The potential value of dowel-nuts was illustrated in construction of a space-frame roof system (Anon. 2000), a demonstration structure (Anon. 2000), and an information kiosk (Green 2001). The author is Professor of Wood Science, Dept. of Forestry and Natural Resources, Purdue Univ., 1200 Forest Products Bldg., West Lafayette, IN This paper was received for publication in July Article No Forest Products Society Member. Forest Products Society Forest Prod. J. 54(12): FOREST PRODUCTS JOURNAL Vol. 54, No

2 In their work, Eckelman and Senft (1995) discussed the potential value of through-bolt with dowel-nut connectors to developing countries in the construction of building frames from small- tree stems. In addition, Stern (2001) noted that The use of small roundwood, especially in less developed parts of the world, will increase for construction as well as for non-structural uses. He went on to say that Effectively connecting small- roundwood is a major problem that must be mastered before small- timber can be used in its most effective manner. Wolfe (2000) outlined the strength advantages of round timber and stated that there is a need to focus research on the development of economically feasible connections to transfer axial loads and bending moment. Initially, it was anticipated that the cost of solid steel dowel-nuts would hamper their use in every day construction and that a modified form of dowelnut would be needed. Furthermore, because considerable care is needed in drilling the holes for the dowel-nut and through-bolt their axes should intersect a connector with less demanding tolerances is advantageous. Finally, dowel-nut construction is limited to those applications where a bolt or threaded rod can be inserted through the hole in the free end of the member and tightened externally from that end. Preliminary tests indicated that through-bolt with cross-pipe and nut connectors in which a short length of pipe is used instead of a solid steel dowel might provide a reasonable alternative to through-bolt with dowel-nut connectors. Presumably, the cross-pipe connectors would be less expensive, have somewhat less demanding tolerances, and allow for assembly and tightening at both ends of the through-bolt (or threaded rod). Through-bolt with dowel-nut or cross-pipe connectors would be expected to be particularly useful for the construction of high-strength T -joints and heel joints of the type shown in Figure 2 in trusses constructed with either round or rectangular members. They should be especially useful as heel joints in joining round or irregular-shaped members together. In addition, through-bolt with cross-pipe connectors provide a means of reinforcing critical joints such as tie beam joints in round mortise and tenon Figure 1. General configuration of through-bolt with dowel-nut specimen used in previous tests by Eckelman and Senft (1995). Figure 2. Possible T -joint and heel joint through-bolt with cross-pipe and nut connection in a truss. joint timber frame construction as shown in Figure 3. They should also be wellsuited for space frame connections as shown in Figure 4. An exploratory test program was undertaken, accordingly, to determine the withdrawal strengths of through-bolt with cross-pipe and nut connections in the ends of selected timbers. Specific objectives of these tests were: a. To obtain estimates of the withdrawal strength of selected cross-pipe, beam size, and species combinations. b. To obtain estimates of the useful and ultimate holding strengths of the cross-pipes. c. To identify and evaluate means of improving the strength and reliability of through-bolt with cross-pipe and nut connections. Results of the study are given in the report that follows. Specimen construction The typical configuration of the tension used in the tests is shown in Figure 5. Two types of were constructed, one with the nut contained inside the cross-pipe (Fig. 5a), and the other with the nut located on the outside of the cross-pipe (Fig. 5b). Figure 3. Through-bolt with crosspipe used to reinforce tie beam connection in round mortise and tenon joint construction. Specimens were constructed of material with rectangular cross sections of approximately 1-1/2 by 3-1/2, 3-1/2 by 3-1/2, 3-3/4 by 3-3/4, and 5-1/4 by 5-1/4 inches. Except where noted, measured 18 inches long. Cross-holes were bored either 5 or 6 inches from each end of a specimen. di- 30 DECEMBER 2004

3 Figure 4. Possible space frame connection. Figure 6. Diagram showing side and end of the specimen reinforced with two 1/4-inch- bolts. Figure 5. Typical configuration of used in the tests. ameters included 1-1/4 (1.660 in. actual outside ), 1-1/2 (1.900 in. outside ), and 2 (2.375-in. outside ) inches. Unless otherwise noted, cross-pipes were fabricated from schedule 80 pipe. Through-bolt (National Coarse [NC] threaded rod) s included 1/2-, 3/4-, and 1-inch. Plain steel threaded rod (NC) was used except in the case of high strength joints constructed with 3/4-inch rod in these, high strength rod was used (B7 alloy steel). Wood species included southern yellow pine (Pinus spp.) and yellow-poplar (Liriodendron tulipifera). The 3-1/2-, 3-3/4-, and 5-1/4-inch square all contained boxed hearts; the 1-1/2 by 3-1/2-inch did not. All material, except where noted, was stored outdoors under roof. Moisture contents (MC), accordingly, were in the 10 to 15 percent range. End holes were drilled with wood bits on a horizontal drill press; cross holes were drilled with Forstner bits on a conventional vertical drill press. End holes were drilled first. These holes were then used as locators in a jig to ensure that the cross holes were drilled concentrically with them. Where noted, the ends of the were reinforced with two 1/4-inch bolts, which were inserted in 5/16-inch relief holes (Fig. 6). Ten were also constructed from salvaged small- yellowpoplar stems. These stems had been cut and piled (with bark) for 18 months prior to use. Specimens 18 inches long were cut from this material (approximately 6 in. in ) and 1-inch holes drilled in each end. The were then stored in an environmental chamber with conditions set to produce an equilibrium moisture content (EMC) of 7 percent, which induced severe splitting (Kubler 1974, 1977). Half of these were reinforced with 1/4-inch- threaded rod hoops as shown in Figure 7. This reinforcement is similar in principle to the hoop and lug system used with silos and water tanks. Hoops were tightened until the yield point of the rod was approached. These were included in the study in order to obtain initial insights into the effectiveness of this type of end reinforcement. Testing procedure All tests were carried out in a universal testing machine. Rate of loading was 0.25 inches per minute. Unless otherwise noted, a test was continued until a major wood failure occurred or until the nut began to pull through the wall of the cross-pipe. Each specimen was photographed during testing. MC samples were cut from the following testing. Results Results for the first set of tests are given in Table 1. All of the yellowpoplar (Yp) and southern yellow pine (SYP) had a cross-pipe to FOREST PRODUCTS JOURNAL Vol. 54, No

4 Table 1. End withdrawal strengths of with through-bolts and 1-1/4-, 1-1/2-, or 2-inch cross-pipes. Nut located inside pipe. Schedule 80 cross-pipes. Wood species a Cross section End distance Cross hole Ultimate load Rod Diameter Length Average SD b (in.) (nom.) (in.) (lb.) (%) 5 SYP 2 by /16 1-1/4 2 1/2 6,088 1, SYP 2 by /16 1-1/2 2 1/2 5,693 1, Yp 3-3/ /16 1-1/4 4 3/4 11,550 2, Yp 3-3/ /16 1-1/2 4 3/4 13,570 4, SYP 3-1/ /16 1-1/4 4 3/4 13,675 4, SYP 3-1/ /16 1-1/2 4 3/4 10,896 2, Yp 5-1/ /16 1-1/2 4 3/4 19,480 3, Yp 5-1/ / ,403 3, SYP 5-1/ / , SYP 5-1/ / ,583 6, a SYP = southern yellow pine; Yp = yellow-poplar. b SD = standard deviation. c COV = coefficient of variation. COV c member end distance of 5 inches except for one set of SYP. Means, standard deviations (SD), and coefficients of variation (COV) are given in Table 1 to allow for a better interpretation of the data. The 1-1/4- and 1-1/2-inch cross-pipes yielded average strengths of 6,088 and 5,693 pounds, respectively, in SYP 2 by 4 s. Likewise, average holding strengths of these pipes in 3-3/4-inch square (in.- sq.) Yp ranged from 11,550 to 13,570 pounds and from 13,675 to 10,896 pounds in 3-1/2 in.-sq. SYP. The 1-1/2-inch pipe had an average holding strength of 19,480 pounds in 5-1/4 in.-sq. Yp, whereas the 2-inch cross-pipes had average holding strengths of 11,170 and 13,403 pounds in 5-1/4 in.-sq. SYP and Yp, respectively. The 2-inch pipes with 6-inch end distance in 5-1/4 in.-sq. SYP had a higher average holding strength of 19,583 pounds. Lowest over highest (L/H) strength ratios varied from 0.41 to a unique high of 0.81 with an average L/H ratio of Observations of the during testing indicated that wood failure arose most often in tension perpendicular to the grain. These failures originated at the end of the member, often beginning at pre-existing surface checks or splits, and then progressing from the end of the member to the cross-pipe (Fig. 8). Results for the second set of tests are given in Table 2. In this set of tests, the ends of the members were reinforced Figure 7. Diagram showing the end of the round small tree specimen reinforced with a hoop and a modified form of lug. with two 1/4-inch- bolts as shown in Figure 6. These bolts were tightened until the washers were embedded to their full thickness in the wood. In addition, the cross hole to member end distance was increased to 6 inches. As can be seen in Table 2, on average, use of the cross bolts substantially increased the strength of the joints. The 1-1/4-inch pipes in 3-1/2 in.-sq. SYP and 3-3/4 in.-sq. Yp averaged slightly less than 20,000 pounds, whereas the 1-1/2-inch pipes in 3-3/4 and 5-1/4 in.- sq. Yp and 5-1/4 in.-sq. SYP averaged well above 20,000 pounds. Likewise, the 2-inch pipes in 5-1/4 in.-sq. SYP averaged slightly less than 30,000 pounds, whereas the 2-inch pipe in 5-1/4 in.-sq. Yp averaged slightly less than 25,000 pounds. Of particular interest with respect to these test results, however, is the substantial decrease in variability of the strength values. This result is reflected Figure 8. Diagram showing tension perpendicular to grain failure at the end of the member. 32 DECEMBER 2004

5 Table 2. End withdrawal strengths of with through-bolts and 1-1/4-, 1-1/2, or 2-inch cross-pipes. Ends of reinforced with 1/4-inch cross bolts. Nut located inside pipe. Schedule 80 cross-pipes. Wood species a Cross section End distance Cross hole Ultimate load Rod Diameter Length Average SD b (in.) (nom.) (in.) (lb.) (%) 5 SYP 3-1/ /16 1-1/4 4 3/4 19,540 1, Yp 3-3/ /16 1-1/4 4 3/4 19,270 1, Yp 3-3/ /16 1-1/2 4 3/4 23, Yp 5-1/ /16 1-1/2 4 3/4 21,980 1, SYP 5-1/ /16 1-1/2 4 3/4 24, SYP 5-1/ / ,650 1, Yp 5-1/ / ,125 1, SYP 5-1/ / /2 d 6 3/4 29,550 2, a SYP = southern yellow pine; Yp = yellow-poplar. b SD = standard deviation. c COV = coefficient of variation. d Indicated 1-1/2-inch pipe nested in a 2-inch pipe. COV c Table 3. End withdrawal strengths of with through-bolts and 1-1/4-, 1-1/2, or 2-inch cross-pipes with the nut located outside of the cross-pipe. Schedule 80 cross-pipes. Wood species a Cross section End distance Cross hole Ultimate load Rod Diameter Length Average SD b (in.) (nom.) (in.) (lb.) (%) 5 SYP 2 by /16 1-1/4 2 1/2 5, Yp 3-3/ /16 1-1/4 4 3/4 11,250 1, SYP 2 by /16 1-1/2 2 1/2 4, Yp 3-3/ /16 1-1/2 4 3/4 8,700 2, Yp 3-3/ /2 d 4 3/4 19, Yp 5-1/ / /2 e ,890 2, a SYP = southern yellow pine; Yp = yellow-poplar. b SD = standard deviation. c COV = coefficient of variation. d Reinforced with cross bolts and two 9/16-inch washers. e Reinforced with cross bolts and two 2- by 1-1/2-inch pipe inserts. COV c in the average of the L/H strength ratios for these, namely, Finally, in one set of tests, a high strength (3/4-in. NC, B7 alloy steel) threaded rod was used with 2-inch pipe that was reinforced internally with a 1-1/2-inch- cross-pipe sleeve. In practice, the through-bolt hole in the sleeve was aligned with the hole in the cross-pipe so that the through-bolt passed through both walls and was anchored with a nut on the inside of the 1-1/2-inch pipe. An average withdrawal strength of 29,550 pounds was obtained with these joints. Minimal damage occurred to the cross-pipes in these tests even at the 33,000-pound load level. Results for the third set of tests are given in Table 3. In these tests, the nut was placed on the outside of the crosspipe (Fig. 5b), so that the cross-pipe was loaded in compression. Failure of these resulted from flattening of the cross-pipe under load. This action caused the cross-pipe to exert pressure on the sides of the cross-pipe hole, which ultimately caused the specimen to fail in tension perpendicular to the grain. In the case of the 1-1/4- by 2-inch schedule 80 cross-pipes in 2 by 4-inch SYP, flattening, i.e., ovalization of the crosspipe, began to be noticeable about 4,000 pounds; specimen failure occurred on average at a little over 5,000 pounds. In the case of the 1-1/2-inch pipe, flattening began at about the same load level; specimen failure averaged about the same as for the 1-1/4-inch pipe. Less flattening occurred with the longer 1-1/4- by 4-inch pipes in 3-3/4 in-sq Yp; failure occurred on average at 11,250 pounds. Flattening of these pipes appeared to commence at about 10,000 pounds. Holding strength of the 1-1/2- inch pipes in the same material averaged less 8,700 pounds owing to two low values. Flattening of the pipes occurred at about 8,000 pounds with this pipe geometry so that values much above 10,000 pounds presumably cannot be obtained without reinforcing the pipe. Following these tests, the cross-pipes were reinforced to resist crushing. Two methods of reinforcement were used. In the first method, two 9/16-inch washers were joined together by means of a number 8 machine screw and inserted inside the 1-1/2-inch cross-pipes, as shown in Figure 9, where they acted as force rings. In practice, it was found that the washers should be positioned as closely as possible to the through-bolt in order to obtain maximum strength. With this reinforcement, an average withdrawal strength of 19,950 pounds was obtained with 1-1/2-inch pipes in 3-3/4 FOREST PRODUCTS JOURNAL Vol. 54, No

6 Table 4. Ultimate pull-through strength of schedule 80 cross-pipes. Pipe Threaded rod Average pull through force SD a (nom.) (in.) (lb.) /4 7 3/4 19,970 1, /2 9 3/4 23, ,600 3,516 a SD = standard deviation. Figure 9. Diagram showing the manner in which two 9/16-inch washers were used to reinforce the 1-1/2-inch pipe against crushing. Figure 10. Diagram showing the 2-inch cross-pipe reinforced with two short lengths of 1-1/2-inch- internal cross-pipes. in.-sq. Yp equipped with 1/4-inch cross bolts. The L/H strength ratio for this set of joints was In the second method, 1-1/2- by 2- inch long lengths of cross-pipe were inserted into the ends of 2- by 6-inch-long cross-pipes (Fig. 10). The 1-1/2-inch pipes were pushed into the ends of the cross-pipe until they butted up against the through-bolt. With this reinforcement, an average withdrawal strength of 22,890 pounds was obtained in 5-1/4 in.-sq. Yp equipped with 1/4-inch cross bolts. The L/H strength ratio for this set of joints was In general, the nuts in all of the tests seated themselves on the bottom of the pipe at load levels about 2,500 to 3,000 pounds less than the ultimate load. At these load levels, significant ovalization of the holes occurred, but no tearing of the metal (associated with pull through) had occurred. The average force required to pull a 3/4-inch nut through the wall of a 1-1/4-inch pipe, Table 4, was 19,970 pounds with a standard deviation of 1,360 pounds. The lowest force required was 18,000 pounds. In the case of the 1-1/2-inch pipe, an average force of 23,878 pounds was obtained with a SD of 975 pounds. The lowest force required was 21,650 pounds. Likewise, the average pull-through force of a 1-inch nut in a 2-inch crosspipe amounted to 32,600 pounds with a SD of 3,516 pounds. The lowest force obtained was 30,000 pounds. It should be noted that pull through occurred in only three. In general, little damage to the pipes was noted below 25,000 pounds, but above this load level, significant ovalization of the holes occurred, which became severe at about 28,000 pounds. Importantly, pull through did not occur in any of the in which a 1-1/2-inch cross-pipe was nested in a 2-inch cross-pipe. The average ultimate strength of the seven strongest of these was 29,129 pounds with a SD of 2,376 pounds. Ovalization of the holes was not severe in any of the pipes. Ultimate pull-through strength, therefore, presumably lies comfortably above 29,000 pounds. The effect of reinforcing round members with hoops as shown in Figure 7 is illustrated in Table 5. As can be seen, reinforcement of the with a hoop of 1/4-inch rod substantially increased the withdrawal strengths of the. Specifically, the withdrawal strength of the without reinforcement averaged 7,238 pounds, whereas the strength of those with reinforcement averaged 17,640 pounds. This result indicates that this form of wrapping is highly effective in reinforcing the ends of circular members. Discussion Initial tests were conducted with a cross-pipe axis to specimen end distance of 5 inches. This relatively short spacing was selected because in practice, joints with small end distances are easier to construct and minimize the cost of threaded rod required, and also because use of a short spacing was thought more likely to produce a recognizable pattern 34 DECEMBER 2004

7 Table 5. End withdrawal strengths of round yellow-poplar without and with hoop reinforcement. Nut located inside cross-pipe. Schedule 80 cross-pipes. End hoop Wood species a Approximate specimen End distance Cross hole Ultimate load Diameter Length Rod Average SD b (in.) (nom.) (in.) (lb.) (%) 4 No Yp Round 6 2-1/ ,238 3, Yes Yp Round 6 2-1/ ,740 3, a Yp = yellow-poplar. b SD = standard deviation. c COV = coefficient of variation. COV c Figure 11b. Diagram showing an alternative manner of creating splits at desired locations on the ends of squared timbers. Figure 11a. Drying splits tend to form on the two sawn faces of round members. This tendency is enhanced if the bark is left on the remainder of the member. of joint failure. Specifically, in previous tests carried out with dowel-nuts in table legs, it was observed that failure often occurred owing to splitting of the end of the leg resulting from tension perpendicular to the grain rather than from shear parallel to the grain of the wood lying beneath the dowel-nut and the end of the specimen. Use of a short spacing helps to prevent masking of this effect by other types of failures and makes the type of failure easier to recognize. It should be noted, however, that use of a short spacing likely reduces the strength values obtained and contributes to their variability. Results of the tests clearly indicate that high strength values can be obtained with through-bolt with cross-pipe and nut connectors in the ends of structural timbers. The tests also indicate that cross-pipes constructed of schedule 80 pipe are relatively well-matched in strength to the strength of the members in which they are embedded. The results obtained can be quite variable, however, and are severely influenced by splits and checks in the members. It was found, however, that high strength values can be obtained in members with a major drying split provided that the longitudinal axis of the cross-pipe is oriented perpendicularly to the plane of the split as shown in Figure 6. This result is particularly important because it was observed during the course of the study that major drying splits could regularly be induced in round members by sawing two flat faces on opposite sides of a member while leaving the bark intact on the remaining surfaces (Fig. 11a). The ends of the members were coated to retard end-drying particularly the area near the pith. Also, holes for throughbolts were not drilled into the ends of the members until they had dried as suggested by Kubler (1974, 1977). It was also found that similar results could be achieved with squared timbers if they were stacked for a short time as shown in Figure 11b. Results of the tests also indicate that the strength of the connections can be increased substantially and the variability in strength greatly decreased by reinforcing the ends of the members. Reinforcement of the ends of the members with two 1/4-inch bolts resulted in strength values of about 20,000 to 25,000 pounds with SDs of about 1,000 to 2,000 pounds. Excellent results also were obtained in roundwood members with a reinforcing rod, or hoop, looped around the end of the member and tightened (Fig. 7). This result agrees with that obtained in earlier tests (Eckelman and Senft 1995) in which the ends of the members were reinforced with 3/4- inch-wide steel strapping. The tests also indicate that the strength of connections with the nut located on the outside surface of the cross-pipe is limited by the crushing strength of the pipe itself, which, as it flattens, presses against the wall of the cross-pipe hole and causes failure. This action severely limits the strength of this FOREST PRODUCTS JOURNAL Vol. 54, No

8 Table 6. Costs of various end connection combinations with 1-foot-long threaded rod based on current retail prices. Add $0.10 to the cost of joints that require 6-inch cross bolts. All costs are in dollars. length Pipe cost/ft. Pipe cost/length Rod Rod cost 1-foot Nut cost Cross bolt length Cross bolt + nut 2 cost Total cost (in.) ($) (in.) ($) (in.) ($) / / / / / / / / / / / / / / / /2 6 4 a High-strength alloy threaded rod / / a type of connection. The strength of these connections was substantially increased, however, when the 1-1/2-inch crosspipes were reinforced with washers used as force rings and when the 2-inch pipes were reinforced with pipe inserts. This result is important because this type of connection can be used in those applications where the nut must be tightened in place. Thus, this joint greatly expands the range of application of cross-pipe connections. Additionally, results of the tests indicate that the cross-pipes themselves behave predictably and are of relatively uniform strength. None of the crosspipes fractured or failed in an unusual manner. Their structural behavior, therefore, should simplify the eventual development of allowable design strength values for these joints. Estimates of the ultimate holding strengths of the crosspipes are provided by the cases in which the nut pulled through the wall of the pipe. Most of these failures occurred in with reinforced ends. Costs of the cross-pipe connections when produced on a large-scale basis are not known; however, the costs of a number of possible end connections based on current retail prices (McMaster-Carr 2002) are given in Table 6. As can be seen, cost per joint ranges from about $1.40 to $5.66. Figure 12. Estimated high, estimated low, and actual ultimate holding strengths for four threaded rod s. Although allowable static design loads cannot be given for the threaded rod included in the study, useful estimates can be obtained from a consideration of their yield strengths. The minimum yield strength of the plain steel threaded rod used in the study was quoted as 58,000 psi with a maximum of 80,000 psi (McMaster-Carr 2002). According to published procedures (Oberg and Jones 1970), the root areas of 1/2-, 5/8-3/4-, and 1-inch threaded rods are 0.142, 0.232, 0.334, and in. 2,respectively. Based on these areas, the corresponding tensile holding capacities of these rods at the minimum yield point amount to 8,230, 13,471, 19,399, and 35,133 pounds, respectively. These values, along with the comparable values for 80,000 psi threaded rod and the ultimate strength values obtained by test are shown in Figure 12. As can be seen, the test values lie above the estimated low yield values but below the estimated high yield values for each rod. 36 DECEMBER 2004

9 If, for the purposes of discussion, allowable static design loads are assumed to be two-thirds of yield point loads (presumably, this value lies below the proportional limits for the rods), the static design loads for these rods are 5,490, 8,981, 12,933, and 23,422 pounds for 1/2-, 5/8-, 3/4-, and 1-inch rods, respectively. Under this assumption, joints based on 1/2-inch rod could be used for axial loads a little over 5,000 pounds with an accompanying cost of $1.40 to $2.12 per connection ($0.37 to $0.42/kip). For loads up to 7,500 pounds and slightly above, 5/8-inch rod could be used with a cost of $1.73 to $2.45 per connection ($0.23 to $0.33/kip). For loads up to 12,500 pounds, 3/4-inch rod could be used with accompanying costs of $2.28 to $3.05 per connection ($0.18 to $0.24/kip). Likewise, for loads to about 22,500 pounds and slightly above, 1-inch rod could be used at a cost of $4.47 per connection ($0.20/kip). Finally, the latter strength also could be obtained with a B7 grade threaded rod and a 1-1/2-inch pipe nested in a 2-inch pipe at a cost of $5.66 per connection ($0.25/kip). Conclusions Results of the tests indicate that high strength end connections can be obtained in structural timbers with throughbolt and cross-pipe and nut connectors. Surface checks and drying splits reduce the strength of the joints and lead to major variations in strength. Proper drying enhances the formation of large drying splits on predetermined surfaces as opposed to multiple smaller random splits and checks. The weakening effect of a major drying split can be reduced by orienting the cross-pipe perpendicular to the plane of the split. Construction of joints with the nut located on the outside as opposed to the inside of the pipe greatly increases the possible uses of the joint. This arrangement provides room for the nut to be tightened in place and allows direct end to end joining of members as well as other constructions that otherwise would not be feasible. Performance of the joints is enhanced greatly when the ends of the member are reinforced, either with some form of strapping or with cross bolts. Not only is the average withdrawal strength increased but the variability in strength is greatly reduced. Acknowledging that there will be exceptions, it can be postulated that use of such reinforcement is almost necessary both because of the high strength values obtained and the reduction obtained in variability of joint strength. The cost of the connectors is expected to be low relative to special fasteners since only a bolt, nut, and short length of pipe are required for construction. Allowable design values for this type of joint need to be developed. Of primary concern is the variability in the strength of the joints. Literature cited Anonymous Forest Products Laboratory research program on small- material. Gen. Tech. Rept. FPL-GTR-110. USDA Forest Serv., Forest Prod. Lab., Madison, WI. 31 pp. Eckelman, C.A Bending strength of plate joints in table construction. Holz als Roh- und Werkstoff. 57(1999): Holding strength of through-bolt with dowel-nut connections. Forest Prod. J. 39(11/12): Evaluating the strength of library chairs and tables. Library Technology Reports. 13(4): [p. 382]. and Y. Z. Erdil Joint design manual for furniture frames constructed of plywood and oriented strand board. FNR-170. Purdue Univ. 34 pp. and J.F. Senft Truss system for developing countries using small roundwood and dowel nut construction. Forest Prod. J. 45(10): Green, S.L. and J. Livingston Exploring the uses for small- trees. Forest Prod. J. 51(9): Kubler, H Formation of checks in tree stems during heating. Forest Prod. J. 27(1): Drying tree disks simply and without defects. Forest Prod. J. 24(7): McMaster-Carr McMaster-Carr Supply Co. Catalog 104. Chicago, IL. Oberg, E. and F.D. Jones Machinery s Handbook, 18th Ed. Industrial Press, NY. 2,293 pp. Stern E.G Construction with small roundwood. Forest Prod. J. 51(4): Wolfe, R Research challenges for structural use of small- round timbers. Forest Prod. J. 50(2):21-29., J. King, and A. Gjinolli Dowel-nut connections in Douglas-fir peeler cores. Res. Paper FPL-RP-586. USDA Forest Serv., Forest Prod. Lab., Madison, WI. 16 pp. FOREST PRODUCTS JOURNAL Vol. 54, No